• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2005년도 추계 학술대회 발표논문집
• /
• pp.163-166
• /
• 2005

We propose a speech emotion recognition method for natural human-robot interface. In the proposed method, emotion is classified into 6 classes: Angry, bored, happy, neutral, sad and surprised. Features for an input utterance are extracted from statistics of phonetic and prosodic information. Phonetic information includes log energy, shimmer, formant frequencies, and Teager energy; Prosodic information includes pitch, jitter, duration, and rate of speech. Finally a patten classifier based on Gaussian support vector machines decides the emotion class of the utterance. We record speech commands and dialogs uttered at 2m away from microphones in 5different directions. Experimental results show that the proposed method yields 59% classification accuracy while human classifiers give about 50%accuracy, which confirms that the proposed method achieves performance comparable to a human.

• 한국음향학회지
• /
• 제16권2호
• /
• pp.41-48
• /
• 1997

본고는 음성인식 기능이 추가된 음성응답장치(ARS)의 음성 인식률을 향상시키는 방법을 제안한다. ARS에 입력되는 전화음성은 안내방송, 전화잡음, 그리고 채널왜곡에 의하여 영향을 받기 때문에, 양질의 음성을 대상으로 하여 개발된 인식 알고리듬을 그대로 적용하면 상당한 인식률의 저하를 가져오게 된다. 이러한 문제점을 극복하기 위하여 본고에서는 세 가지 방법을 제안한다: 1)음성이 시작되는 순간 안내 방송을 즉시 끊기 위한 음성 입력순간의 정확한 검출, 2)Teager 에너지를 이용한 잡음 섞인 전화음성의 효과적인 끝점검출, 3)SDCN 알고리듬을 이용한 채널왜곡의 보상. 위의 세 가지 방법을 모두 결합하여 화자독립인 전화음성을 대상으로 실험한 결과, 기존의 방법이 약 23%의 인식률을 보인 반면, 제안된 방식은 약 77%의 인식률로서 상당한 성능향상을 보여주었다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1996년도 춘계학술대회 논문집
• /
• pp.259-263
• /
• 1996

In diagnosing of mechanical machinery, it is often improtant to get information about the movement inside the machine casing. If the values of internal tities may be derived from the measurement using sensors installed on the external casing, it would be much better in many senses. This paper discusses extracting internal gear movements byfrequencydemodulation from gear meshing force signatures which can be recovered from the vibrations though inverse filter. There are several way in demodulating signals. In this paper, especially, Hibert Transform, Wigner-Ville distribution, and Teager energy operator are examined and compared. Effects of noise on the frequency demodulation methods and the behavior of bandpass filtered noisy signal are discussed using simulated time-varying frequency signals.

• International Journal of Computer Science & Network Security
• /
• 제23권8호
• /
• pp.40-48
• /
• 2023

Speech Emotions recognition has become the active research theme in speech processing and in applications based on human-machine interaction. In this work, our system is a two-stage approach, namely feature extraction and classification engine. Firstly, two sets of feature are investigated which are: the first one is extracting only 13 Mel-frequency Cepstral Coefficient (MFCC) from emotional speech samples and the second one is applying features fusions between the three features: Zero Crossing Rate (ZCR), Teager Energy Operator (TEO), and Harmonic to Noise Rate (HNR) and MFCC features. Secondly, we use two types of classification techniques which are: the Support Vector Machines (SVM) and the k-Nearest Neighbor (k-NN) to show the performance between them. Besides that, we investigate the importance of the recent advances in machine learning including the deep kernel learning. A large set of experiments are conducted on Surrey Audio-Visual Expressed Emotion (SAVEE) dataset for seven emotions. The results of our experiments showed given good accuracy compared with the previous studies.

• 전기학회논문지
• /
• 제67권7호
• /
• pp.816-823
• /
• 2018

The Auto-synchronizer is essential equipment for synchronizing a generator to the power system. It is performing that measurement of the magnitude, frequency and phase of the voltage signal of the power system and generator. It is important to select the appropriate measurement algorithm for preventing various problem such as mechanical stress and Electrical problem. Teager Energy Operator(TEO) and Discrete separation algorithm(DESA) is measurable the instantaneous parameters of a sine wave using 5 samples and can be measured at a fast and with a simple operation. Therefore it has many advantages in measuring the parameters. In this paper, it confirmed measurement results using matlab simulations when there are synchronized in order of frequency, magnitude. Also it presented methods using digital filters and sample intervals to improve accuracy.